Conventional optimization models for electric vehicle (EV) charging do not sufficiently account for battery wear. Accordingly, disclosed embodiments utilize a heuristic approach to penalize optimization solutions for deviations of electric vehicles or other flexible loads (e.g., microgrid batteries) from state-of-health (SoH) parameters. For example, the SoH parameters may encourage the optimization model to quickly bring and maintain states of charge for the batteries of the flexible loads within an ideal state-of-charge (SoC) range. This approach biases the optimization model towards a solution that reduces the occurrence of such deviations from SoH parameters, thereby extending the lifetime of the batteries.